Open Access

MLPA-conservation mechanism in wireless sensor network environments

  • Sheng-Tzong Cheng1,
  • Jia-Shing Shih1Email author,
  • Tun-Yu Chang1,
  • Gwo-Jiun Horng1 and
  • Chih-Lun Chou1
EURASIP Journal on Wireless Communications and Networking20122012:343

https://doi.org/10.1186/1687-1499-2012-343

Received: 16 December 2011

Accepted: 4 October 2012

Published: 16 November 2012

Abstract

Transmission using excess power not only shortens the lifetime of sensor nodes, but also introduces interference. Packets should ideally be transmitted with moderate power. This study proposes a multi-level power adjustment (MLPA) mechanism for a wireless sensor network to prolong the lifetime of individual nodes and the overall network. In this study, we constructed an analytical model of the MLPA mechanism with m distinct power levels (m-LPA). For m-LPA, the closed-form expression of the optimal power setting was determined and the mean transmission power was minimized to one-third of the original fixed-transmission power. We found that the average power consumption of our proposed mechanism is 33.93% higher than that of fixed-transmission power. Thus, each node can extend the lifetime by 2.5 times. We have shown the relations between m and density in simulation results. Although the mechanism worked smoothly in this study, the sensors do not need to handle the distance and interference problem.

Keywords

Lifetime Wireless sensor network Power adjustment Transmission power Interference

Introduction

Power consumption has become one of the main topics in research on wireless sensor networks (WSNs). In the past few years, there has been much improvement in the technology and the application of WSNs [13]. Researchers have compiled complete sets of standards for real cases (e.g., IEEE 802.15.4 [4] and Zigbee) spanning a wide range of applications including home security, private sector organizations, health care, and the military. In general, each sensor has limited power and computing resources. Therefore, a major issue is the prolongation of network lifetimes for WSNs.

There are three basic subsystems on a sensor node: a sensing subsystem to collect data from the environment, a processing subsystem to provide computation ability to the sensor node, and a communication subsystem for the transmission of data [5]. These subsystems are all limited by internal sensor-node power. Each sensor node plays its part in performing sensing, computing, or transmitting, always within the context of limited energy. In order to extend the lifetime of WSNs, we focus our attention to transmission problems. Theoretically, transmission power can be separated into n portions. In this study, we define the level is n portions of transmission power. Each sensor must identify the number of power-adjustment level. And each sensor cans effective delivery packets to next node. In addition, each node should save transmission power efficiently. We propose a power-control mechanism that achieves efficient sensor-data transmission resting on two special features: (1) an adaptive transmission power rate for packet routing and (2) a prolongation of sensors’ lifetime.

The major contributions of this article are as follows.

  • We provide an multi-level power adjustment (MLPA) mechanism for implementation in a real WSN environment, and we prolong the lifetime of WSNs.

  • We present a rigorous analysis of an MLPA mechanism and identify near-optimal power configurations for each power level.

  • We present our findings that our proposed mechanism’s (2LPA, 200–500 nodes) average power consumption registers a 33.93–55.30% improvement over the original fixed-transmission power method. Thus, the lifetime of each node can extend 2.5 times.

  • As the simulation results show, we found that when we adjusted the power level to 2 or 3, we could get acceptable performance. If the power level was more than 3, the lifetime performance was better.

  • The mechanism that we proposed can work smoothly even though to get the distance value between nodes and node is insufficient and even though there is some interference between nodes. In other words, the proposed mechanism can still work even if the between-node distance value is unsatisfactory, the sensor board lacks GPS chips, or the environment exhibits refraction and attenuation.

Related work

This section describes the literature’s treatment of WSNs and various approaches to determining the transmission ranges for the extension of WSN lifetimes [610]. Conventional WSNs consist of thousands of sensor nodes and sensors that can communicate with each other or with base stations. The coverage area of a WSN depends on the number of sensors deployed [1113].

Every sensor node has sensing, processing, transmitting, moving, locating, and electrical power elements; some elements are optional, such as the moving or locating element. Sensor nodes are placed in a particular area and integrated with each other to yield high-quality information about the environment [14]. Every sensor node not only can capture the sensed data and route it to other sensors or base stations, but also can determine methods of routing or scheduling according to the mission, the received data, and the status of the sensor node. A base station is a mobile or static device for connecting a WSN to the Internet, thereby enabling a user to receive sensed data.

In recent years, considerable research has addressed the coordination and management of sensing, as well as sensor-related data collection and processing [1]. Due to the limitations of both electrical power and sensor nodes’ communication bandwidth, it is necessary to design a technology that reduces the waste of power and that employs limited bandwidth effectively. The limitations in traditional WSNs create many challenges of design and management such as electric power’s relationship to each layer of a network’s protocol stack. For example, it is necessary for the network layer to use a power-saving routing method that sends data from sensor nodes to the base station effectively and that increases the network’s lifetime satisfactorily.

The routing methods of current WSNs include table-driven, demand-driven, and hybrid methods. A detailed explanation of these three methods is given below.
  1. (1)

    Table driven method: According to the routing table in every sensor node, the sensed data can be transmitted by the default routing path in the routing table such as destination sequenced distance-vector (DSDV) [15], CGSR [16], and WRP [17].

     
  2. (2)

    Demand driven method: This method does not call for a default routing path. Instead, there is an on-demand path, which is found when source nodes must transmit data to a destination node. The routing path is established through flooding that involves node-path information from the source to the destination. AODV [18] and DSR [19] are examples of this kind of routing method.

     
  3. (3)

    Hybrid method: This method is a combination of the table-driven and demand-driven methods. An example is ZRP [20].

     

Whereas at least three methods for solving problems of data transmission and of electrical power are based on routing, some approaches solve such problems by exploiting communication among sensor nodes [610, 21, 22]. These approaches, when successful, adjust communication power according to the between-node distance values [10], but they must solve problems concerning unknown transmission distance, uncertain range of sensor-node coverage, and nonlinear deployment in real WSNs.

Dhawan et al. [23] proposed a smoothly varying range model to adjust sensing range sensor networks for maximizing the network lifetime. Authors use the Garg-Könemann algorithm 1 with an approximation ratio to find a monitoring schedule for sensor covers. However, the continuous model needs complicate analysis. Wang and Medidi [24] proposed local sensing radii optimization schemes to minimize the energy consumption and extend the lifetime of networks. Authors assume that the transmission range is at least two times larger than the maximum sensing range, and then focus on optimizing sensing coverage for reliable surveillance and energy-efficiency. However, this method requires additional mechanisms and overhead to ensure the connectivity of entire network if obstacles exist between sensors. In [21, 22], the authors divide the sensing field into a number of smaller areas. Based on those smaller area, sensors could control and minimizing energy consumption.

Because of these uncertainties and other unknowns, real WSNs cannot easily compute the precise communication power, even though some investigations have presented corresponding algorithms. This study proposes a power-adjustable routing policy, in which no sensor node transmits data by static communication power. Rather, each sensor node transmits data using dynamic communication power, determined by the conditions of the immediate environment.

Multi-level power-adjustment mechanism

In general, the required transmission power is the same for the communication between all sensor nodes and their neighboring nodes. Consider the example of the WSN shown in Figure 1. Regarding two pairs of sensor nodes, the distance between the first pair’s two members differs from the distance between the second pair’s two members, but the transmission power that consumes in sending data from one pair’s member to the same pair’s other member is the same for each pair. Consequently, the energy efficiency is unsatisfactory, because even if two sensor nodes are so close to each other that data can be transmitted using half the communication power (such as nodes c and d in Figure 1), the traditional method still transmits using the full communication power.
Figure 1

The situation of data-transmission power in traditional WSNs.

Adjustable transmission power is not a new concept. Lin et al. [25] presented adaptive transmission power control for WSNs, where each node constructs a model for each of its neighboring nodes that describe the relationship between transmission power and link quality. A feedback-based transmission power control algorithm serves to maintain individual link quality over time dynamically.

This study presents a power-adjustment table (PAT) for every sensor node to determine the optimal communication power dynamically. Using the PAT, every sender node determines the communication power necessary for transmitting data and does so according to the environment of the nearby receiver node, the objective being to reduce power consumption. This method, accordingly, extends the WSN lifetime. After the adaptive communication power is determined (as in Figure 2), each sender node uses a particular communication power to transmit data consistent with the environment of the nearby receiver node.
Figure 2

The data transmission after determination of the adaptive communication power.

In the proposed MLPA mechanism, the transmission corresponding to each sensor node involves adjustable power. The key feature of MLPA is the use of just enough power to communicate with neighboring nodes. Less power is used for closer nodes. The MLPA mechanism operates according to the following three phases.

  • Phase I: Neighbor discovery and construction of network topology

    Initially, each node broadcasts a beacon to identify all possible neighboring nodes. Full transmission power functions to maximize the network connectivity level.

  • Phase II: Negotiation and adjustment of transmission power negotiation

    Energy is conserved by reducing the average transmission power. Full power is not necessary for transmission to a nearby neighbor. In this phase, a sensor node negotiates with its neighboring nodes and determines which power level should correspond to each neighbor. Given the negotiation overhead, the piggyback mechanism can function to reduce the overhead involved in power negotiation.

  • Phase III: Runtime maintenance

    The communication environment may change dynamically after node deployment. In addition, the movement of nodes may change network topology. Sensor nodes must repeat phases I and II to update the network topology and preserve the optimal power assignment for each neighbor.

    We use MLPA mechanisms to improve energy conservation. Our mechanism and analysis can be divided into two parts: (1) multi-level construction of a PAT; (2) optimal multi-level power-configuration analysis.

Multi-level construction of a PAT

For this study, we initially assumed that each sensor node would have the ability to adjust its own power. Theoretically, transmission power can be separated into n portions, and each sensor must identify the number of power-adjustment level. And each sensor scans effective delivery packets to next node. In addition, each node should save transmission power efficiently. There are some solutions which will precisely compute transmission power by distance. The sensors always use maximum transmission power to deliver packets because sensors have no internal GPS chip for computing the distance between each node and because real environments involve significant interference and attenuation. In this study, we propose a mechanism to solve this problem. We assume that transmission power is separated into three portions: L, M, and S. We use this assumed scenario to explain MLPA construction. Figure 3 shows the coverage of these three portions of transmission power.
Figure 3

Different transmission powers with different signal coverage.

In the WSN’s initial stage, each sensor node will transmit on the basis of these three power discover messages in order to collect neighboring nodes’ information. When we use transmission power L to transmit a discover message, we collect neighboring nodes’ information in power L coverage area. And when we use transmission power M to transmit a discover message, we collect neighboring nodes’ information in power M coverage area. This is the pattern that repeats itself. Each sensor node transmits these three powers in the WSN’s initial stage and, if successful, senses other neighbors’ sensors in its coverage area. These data are stored in a PAT, as given below.

The PAT is stored in the MAC layer. When the routing protocol goes into operation to determine the next transfer node of a packet, our proposed mechanism picks an appropriate power level for this packet in the MAC layer. This study transmits packets by the transmission power level in the physical layer. Through this PAT, we can determine a node’s neighboring node, transmission power consumption, the operating (alive) status of the node, and the response status of the node. A sensor node can optimize transmission power using this PAT to complete the transmission.

As shown in Figure 4, node i transmits data through node k to node j. Table 1 presents our assumed scenario involving a PAT for node k. When node k receives data from node i, node k already knows that the next node is node j. Hence, node k uses level-M transmission power to transmit data to node j, after querying the PAT for node k.
Figure 4

Use of appropriate transmission power through a PAT.

Table 1

Example of a PAT in a sensor node

Neighboring node

Transmitted power

Alive

Response

a

L

1

1

j

M

1

4

e

M

0 (dead)

(null)

g

M

2 (asleep)

3

f

M

1

4

i

S

1

1

When a WSN is deployed in a real environment, measuring the distance between nodes is difficult. Even if the distance between sensor nodes is known, avoiding interference and the effects of nearby obstacles is difficult. Accordingly, using mathematical analysis to calculate transmission power accurately is almost impossible. However, the mechanism proposed herein can take into account environmental conditions between sensor nodes. Consequently, every sensor node can transmit data with optimal power.

Optimal multi-level power-configuration analysis

Our analysis reveals that a WSN can reduce the energy consumed by transmission through an adaptive power-routing strategy. The ratio of retrenched energy consumption is based on the levels of transmission power. These affect the energy efficiency in a WSN. This section will analyze the specific value of transmission power in order to determine the optimal levels of transmission power.

Each sensor node differently adjusts the transmission power. Multiple transmission power adjustments must be made in the situation represented by Figure 5. Theoretically, transmission power can be cut into m-LPA. This study discusses the example of a 3-LPA power-adjustment method, and makes the following assumptions.
  1. (1)

    N sensor nodes are randomly deployed over an area, A. The node density (ρ) is defined as ρ = N/A.

     
  2. (2)

    For an arbitrary node, the transmission probabilities to all neighboring nodes are equal.

     
  3. (3)
    Assume that a signal is attenuated by distance d raised to the power β, where β is the path-loss exponent (PLE). The power of the intended signal from the transmitter at the receiver can be written as [26]
    P r = α P t d β ,
    (1)
    α = G t G r λ 2 4 π 2
    (2)
     
where P t stands for the transmission power at the transmitting antenna, P r denotes the signal power measured at the receiving antenna, G t and G r denote the transmitter and receiver antenna gains, and β denotes the wavelength of the carrier. The PLE depends on the wireless environment and typically varies from two to four. Initially in the free space loss (FSL) model, PLE β = 2 is considered.
Figure 5

MLPA.

The largest transmission power, P, is determined. Then, the maximum transmission radius R can be determined. Without any power adjustment, the sensor nodes transmit data using P, in line with the above statements. Various transmission radii can be derived, r1R, r2R, …, rm–1R, r m R, where 0 < r1 < r2 < ··· < rm–1 < r m = 1, for a sensor node that makes an m-level transmission power adjustment, and the relation between transmission power and transmission distance can be obtained through an FSL model.
P i = r i 2 P , i = 1 , 2 , , m .
(3)
When a sensor node transmits data using transmission power P i , the sensor nodes at distances between ri-1R and riR can receive the data. Therefore, the number of sensor nodes in the range is
n i = ρ π r i 2 r i 1 2 R 2 , for i = 2 , 3 , , m .
(4)
All sensor nodes are assumed to have equal transmission probabilities and access. Therefore, the equation that governs the m-level adjustment at the average power consumption is as follows.
P a v g m = i = 1 m P i n i i = 1 m n i .
(5)
The 3-LPA power adjustment proposed in this study yields three transmission ranges.
r 1 R , r 2 R , and r 3 R , where 0 < r 1 < r 2 < r 3 = 1 .
(6)
Transmissions in the three ranges are received according to the following transmission power ratios.
P 1 = r 1 2 P
(7)
P 2 = r 2 2 P
(8)
P 3 = P
(9)
The above ratios and distances yield the number of nodes in each transmission range.
n 1 = ρ π r 1 R 2
(10)
n 2 = ρ π r 2 R 2 ρ π r 1 R 2 = ρ π R 2 r 2 2 r 1 2
(11)
n 3 = ρ π R 2 ρ π r 2 R 2 = ρ π R 2 1 r 2 2
(12)
Then, the average transmission power consumption is calculated as follows.
P a v g 3 = P 1 · n 1 + P 2 · n 2 + P 3 · n 3 n 1 + n 2 + n 3 = r 1 4 + r 2 2 r 1 2 r 2 2 + 1 r 2 2 P
(13)

Therefore, to minimize the average transmission power, not only must one minimize f(r1, r2), but also both r1 and r2 must satisfy 0 < r1 < r2 < r3 = 1.

Figure 6, produced in MATLAB, is a three-dimensional representation of the transmission power ratio for r1, r2, and the average transmission power consumed.
Figure 6

A diagram of the transmission power ratio and the average transmission power consumption relative to 3-LPA.

On the basis of the above figure, one can find an ideal power ratio by using differential calculus to minimize 3-LPA:
f r 1 r 1 , r 2 = f r 1 , r 2 r 1 = 4 r 1 3 2 r 1 r 2 2
(14)
f r 2 r 1 , r 2 = f r 1 , r 2 r 2 = 2 r 2 + 4 r 2 3 2 r 2 r 1 2
(15)
The power ratios at r1 and r2 are obtained as follows.
{ 2 r 2 2 r 1 + 4 r 1 3 = 0 2 r 2 + 4 r 2 3 2 r 2 r 1 2 = 0
(16)
r 1 = 3 3 r 2 = 6 3
Therefore, the 3-LPA power ratio for L:M:S is 1 : 6 3 : 3 3 . These calculations enable one to obtain the average transmission power consumption Pavgmin(3) = (2/3)P at r 1 = 3 / 3 and r 2 = 6 / 3 . The minimum average transmission power is (2/3) P for 3-LPA in an FSL environment. The optimal configuration is as follows.
{ P 1 = 1 / 3 P P 2 = 2 / 3 P P 3 = P , { R 1 = 3 / 3 R R 2 = 6 / 3 R R 3 = R , a n d { n 1 = ρ π R 2 / 3 n 2 = ρ π R 2 / 3 n 3 = ρ π R 2 / 3
(17)

The above analysis reveals that appropriately setting the specific value of transmitter power can reduce the transmitter power to a level that is one-third less than that of the fixed-transmission power method.

Optimal power configuration

For the case of PLE β ≠ 2, the power configuration of each power level corresponds to a closed-form expression. The minimum average transmission power of m-LPA is given by the following recursive formula.
P avg min m , β = P m = 1 P β P β + 2 × 2 P β + 2 P avg min m 1 , β 2 / β m = 2 , 3 , 4…
(18)
The recursive formula can be verified by mathematical induction. The power configuration of each level depends on the value of r i , which can be described as
r i = { j = i m 1 q j 1 i < m 1 i = m
(19)
where q j = r j r j + 1 = 2 P β + 2 P avg min j , β 1 / β . When the value of r i is computed, the transmission power of sensor can be computed as follows.
P i = r i β P .
(20)
R i = r i R ' ,
(21)

where R' = R2/β is the maximum transmission range at power P in the wireless environment of PLE β.

Figure 7 plots the minimum average transmission power versus the degree of MLPA. The number of power levels varies from 1 to 100. At m = 100, the minimum average transmission powers are 0.671, 0.505, 0.405, 0.338, and 0.29 for PLE β = 1, 2, 3, 4, and 5, respectively.
Figure 7

Minimum average transmission power for m-LPA.

Near-optimal power configuration

As the degree of m-LPA increases, manually optimizing the configuration grows difficult. The built-in function “Minimize” in Mathematica [27], a powerful scientific and mathematical software package, thus serves to minimize Pavg(m), subject to the constraint 0 < r1 < r2 < ··· < r m = 1. Table 2 presents the optimal configuration and the minimum average transmission power for m-LPA, as m varies from two to five.
Table 2

Optimal configuration for m-LPA

 

r 1

r 2

r 3

r 4

r 5

P avg min(m)

m = 2

1 2

1

3 4 P

m = 3

1 3

2 3

1

2 3 P

m = 4

1 4

1 2

3 4

1

5 8 P

m = 5

1 5

2 5

3 5

4 5

1

3 5 P

Given Table 2, we suggest a closed-form expression for the optimal configuration for m-LPA, where m is a positive integer. In the optimal power configuration of m-LPA (which minimizes the average transmission power), the r i configurations are conjectured to be
r i = i / m .
(22)
Consequently, each transmission power is described as
P i = i m P .
(23)
When m-LPA is used, the average transmission power can be minimized to
P avg min m = m + 1 2 m P .
(24)
Figure 8 plots the minimum average transmission power versus the degree of MLPA (m). As shown in Figure 8, the minimum average transmission power is a monotonically decreasing function of m. Equation ([24]) reveals that a higher degree of MLPA corresponds to a greater transmission power saving. The limit of P avg min(m) P is reached as m approaches infinity.
lim m P avg min m = P 2 .
(25)
Figure 8

Minimum transmission power for m-LPA (β = 2).

Ideally, half of the transmission power can be saved as the number of power levels approaches infinity if the optimal power is used for transmission to each neighboring node. However, providing an MLPA mechanism with an infinite number of power levels is both unfeasible and unnecessary. The power negotiation overhead must also be considered. For m = 10, the maximum transmission-power savings are 45%, which is 90% of the maximum possible power savings in the FSL model. In general, the use of optimized m-LPA yields a transmission power savings of 100 m - 1 2 m %, which is 100 m - 1 m % of the maximum possible power savings in the FSL model.

Performance analysis and results

We wrote a simulation program to confirm the MLPA mechanism; and we compared static transmission power and transmission under the MLPA mechanism. Adjusting transmission power according to the proposed mechanism enables the packets to arrive at the sink node, but reduces the power consumption by almost half.

When the MLPA mechanism is applied, the transmission power undergoes adaptive changes to deliver packets to the next node. The routing protocol is assumed to be table-driven. When routing data, the sensor node looks up the routing table. In the simulation, DSDV [15] is used as the routing protocol. We compared the proposed mechanism to the one that uses fixed transmission power (Fixed-TP).

To determine the efficiency of the MLPA mechanism, we performed a simulation using Visual C++. The sensor area was set to 600 × 600 m2. The number of deployed sensor nodes ranged between 200 and 900. The transmission range of the sensor was 50 m. When the simulation ran, if the number of sensor nodes was under 200, the sensor network could not maintain its network connectivity and WSN failed to cover some regions. Therefore, at least 200 sensor nodes are needed to achieve adequate network functioning.

The DSDV was used as the routing protocol. DSDV is an enhanced version of the distributed Bellman-Ford algorithm. In DSDV, each node maintains the routing table that includes (1) the shortest distance between nodes and sink and (2) the first node on the shortest path to every other node in the network. DSDV includes table updates with increasing numbers of sequence tags to prevent loops.

In this article, we refer to T-mote sensors set to specific power parameters. Simulation program assigns the packet-generation probability in accordance with Poisson probability. In the simulation, we compared the MLPA mechanism (m = 2 to 9) with the traditional Fixed-TP mechanism, which uses only a simple kind of power to deliver information (m = 1). Figure 9 shows that we saved an average of more than 47% of the average power consumption regarding the low-density group (number of nodes between 200 and 500), and Figure 10 shows that we saved an average of nearly 62% of the average power consumption regarding the high-density group (number of nodes between 600 and 900).
Figure 9

Average power consumption for m-LPA (200500 nodes).

Figure 10

Average power consumption for m-LPA (600900 nodes).

Figure 9 shows the average power consumption of WSNs when (1) the effect of m from 2 to 6 is greater than that of m on over 6 and (2) when 2 < m < 6. The comparison between the scheme with Fixed-TP and MLPA scheme with 2 LPA shows that our scheme saves 33.93–55.30% energy under 200–500 nodes in Figure 9. The average power consumption saves 47.04% under sparse environment in summary. Figure 10 shows the average power consumption for efforts to strengthen the effect of m from 2 to 3. The average power consumption saves 61.92% under density environment (600–900 nodes). Furthermore, transmission energy consumption can be saved up to 69.61% using MLPA scheme with 9 LPA.

Figure 11 shows the m-LPA compared with the Fixed-TP regarding average power consumption. A sensor node’s average power consumption for Fixed-TP was 0.01436 mJ in a 200-node environment, but in the same environment, the 2-LPA set up needed just 0.009487 mJ; for the 9-LPA set up, the amount needed was only 0.004338 mJ. Our proposed method is better than Fixed-TP when the objective is to significantly lower the average power consumption.
Figure 11

Average power consumption of sensor node with m-LPA and Fixed-TP.

Figure 12 shows the lifetimes of WSN with m-LPA, and the figure compares our proposed mechanism with Fixed-TP. Fixed-TP’s use of full power to transmit every packet will result in a shortened lifetime for WSNs. The MLPA method allows for the selection of different types of transmission power to deliver packets, and in this regard, Figure 12 presents some interesting findings. MLPA’s lifetime turned out to be much longer than Fixed-TP’s lifetime.
Figure 12

Lifetime(s) of WSN with m-LPA and Fixed-TP.

Figures 13 and 14 show that the Fixed-TP lifetime are short. We compared the lifetimes of the MLPA mechanism with those of the 2-LPA and Fixed-TP approaches, and found that 2-LPA had an average extension of 2.5 times the lifetimes of WSNs. In Figure 13, we can see that when the density of deployment is low, lifetimes under the MLPA mechanism (2 < m 9) can be extended. And Figure 14 shows that when m = 2 and 3, lifetimes get a boost. The growth in lifetimes is less pronounced than the growth when m > 3.
Figure 13

Lifetime(s) of WSN with m-LPA (200500 nodes).

Figure 14

Life time of WSN with m-LPA (600900 nodes).

From Figure 15, we can also understand the effect that density has on lifetimes. The Fixed-TP performance is far below the m-LPA (m > 1) performance. And when the density is less than 0.17 and m > 4, lifetime performance is quite remarkable, but when the density is greater than 0.17 and m > 4, the lifetime curve is flat. For different density environments, generally increasing the m value will enhance and improve the lifetime of whole WSNs. Figure 15 shows that performance tends to get overlap when m = 8 and 9 in situations where the density is greater than 0.17.
Figure 15

Relationship between lifetimes of WSN and density.

Therefore, any effort to set the m value in relation to the environment must account for the density of deployment. In general, m = 2 or 3 can yield a pretty good performance.

Conclusion

We proposed an MLPA mechanism for implementation in real WSN environments to prolong the lifetime of WSNs. And we present a rigorous analysis of an MLPA mechanism and identify near-optimal power configurations for each power level. On the basis of PAT construction, appropriate transmission power can be adjusted during data transmission. We present our findings that the average power consumption of our proposed mechanism (2LPA, 200–500 nodes) registers a 33.93–55.30% improvement over the original Fixed-TP method. Thus, the lifetime of each node can extend 2.5 times. And we have presented the relations between m and density in simulation results. Setting the appropriate m value will improve lifetime performance.

As the simulation results show, we found that when we adjusted the power level to 2 or 3, we could get acceptable performance. If the power level was more than 3, the lifetime performance was better.

The mechanism that we proposed can work smoothly even though to get the distance value between nodes and node is insufficient and even though there is some interference between nodes. In other words, the proposed mechanism can still work even if the between-node distance value is unsatisfactory, the sensor board lacks GPS chips, or the environment exhibits refraction and attenuation.

We conclude that the proposed mechanism can work smoothly. If those sensors cannot get the distance between nodes and there is some interference between nodes, our proposed mechanism can work smoothly.

Declarations

Authors’ Affiliations

(1)
Department of Computer Science and Information Engineering, National Cheng Kung University

References

  1. Ye W, Heidemann J, Estrin D: Medium access control with coordinated adaptive sleeping for wireless sensor networks. IEEE/ACM Trans. Netw. 2004, 12(3):453-506.View ArticleGoogle Scholar
  2. van Dam T, Langendoen K: An adaptive energy-efficient MAC protocol for wireless sensor networks. 1st ACM Conference on Embedded Networked Sensor Systems(ACM SenSys) ed. by (Los Angeles, CA, 2003), pp. 171–181Google Scholar
  3. Tilak S, Abu-Ghazaleh NB, Heinzelman W: Infrastructure tradeoffs for sensor networks. Proceedings of the 1st ACM International Workshop on Sensor Networks and Applications (WSNA) ed. by (Atlanta, 2002), pp. 49–58Google Scholar
  4. IEEE Std 802.15.4, Part 15.4: Wireless medium access control (MAC) and physical layer (PHY) specification for low rate wireless personal area networks (LR-WPANs). 2003.Google Scholar
  5. Santos RA, Villasenor L, Gallardo J, Edwards A: Performance analysis of routing strategies for wireless sensor networks. IFIP 2007, 248: 191-202.Google Scholar
  6. Althaus E, Calinescu G, Mandoiu II, Prasad S, Tchervenski N, Zelikovsky A: Power efficient range assignment for symmetric connectivity in static ad hoc wireless networks. Wirel. Netw 2006, 12: 287-299. 10.1007/s11276-005-5275-xView ArticleGoogle Scholar
  7. Huang M, Caines PE, Malhamé RP: Uplink power adjustment in wireless communication systems: a stochastic control analysis. IEEE Trans. Automatic Control 2004, 49(10):1693-1708. 10.1109/TAC.2004.835388View ArticleMathSciNetGoogle Scholar
  8. Panichpapiboon S, Ferrari G, Tonguz OK: Optimal transmit power in wireless sensor networks. IEEE Trans. Mob. Comput. 2006, 5: 1432-1447.View ArticleGoogle Scholar
  9. Ramanathan R, Rosales-Hain R: Topology control of multihop wireless networks using transmit power adjustment. Proc. IEEE Infocom’2000 ed. by (Tel Aviv, Israel, March 2000), pp. 404–413Google Scholar
  10. Gao Q, Blow KJ, Holding DJ, Marshall IW, Peng XH: Radio range adjustment for energy efficient wireless sensor networks. Ad Hoc Netw. 2006, 4(1):75-82. 10.1016/j.adhoc.2004.04.007View ArticleGoogle Scholar
  11. Akyildiz IF, Su W, Sankarasubmmaiam Y, Cayirci E: A survey on sensor networks. IEEE Commun. Mag. 2002, 40(8):102-114. 10.1109/MCOM.2002.1024422View ArticleGoogle Scholar
  12. Al-Karaki JN, Kamal AE: Routing techniques in wireless sensor networks: a survey. IEEE Wirel. Commun. 2004, 11: 6-28.View ArticleGoogle Scholar
  13. Intanagonwiwat C, Estrin D, Govindan R, Heideman J: Impact of network density on data aggregation in wireless sensor networks. International Conference on Distributed Computing Systems ed. by (Vienna, 2002), pp. 457–458Google Scholar
  14. Gao Y, Wu K, Li F: Analysis on the redundancy of wireless sensor networks. Proc. 2nd ACM International Workshop on Wireless Sensor Networks and Applications (WSNA) ed. by (San Diego, 2003), pp. 108–114Google Scholar
  15. Perkins CE, Bhagwat P: Highly dynamic destination-sequenced distance-vector routing (DSDV) for mobile computers. Comput. Commun. Rev. 1994, 24(4):234-244. 10.1145/190809.190336View ArticleGoogle Scholar
  16. Chiang CC, Wu HK, Liu W, Gerla M: Routing in clustered multihop, mobile wireless networks. Proc. of IEEE Singapore International Conference on Networks (SICON’97) ed. by (Singapore, Apr. 14–17, 1997), pp. 197–211Google Scholar
  17. Murthy S, Garcia-Luna-Aceves JJ: A routing protocol for packet radio networks. ACM/IEEE MOBICOM ed. by (Berkeley, CA, Nov. 1995), pp. 86–95Google Scholar
  18. Perkins CE: Ad hoc on demand distance vector (AODV) routing. IETF Internet Draft (November 1998).[http://www.ietf.org/internet-drafts/draft-ietf-manet-aodv-02.txt]
  19. Johnson DB, Maltz DA: Dynamic source routing in ad-hoc wireless networks. Mob. Comput. Chapter 1996, 5: 153-181.View ArticleGoogle Scholar
  20. Hass ZJ: The zone routing protocol (ZRP) for ad hoc networks, IETF, internet draft, draft-zone-routing-protocol-00.txt. 1997.Google Scholar
  21. Nickray M, Afzali-Kusha A, Jäntti R: MEA: an energy efficient algorithm for dense sector-based wireless sensor networks. EURASIP J. Wirel. Commun. Netw. 2012.Google Scholar
  22. Mundra PS, Kumar R: Energy efficient protocol for WSN. Int. J. Eng. Res. Appl. (IJERA) 2012, 2(3):2912-2916.Google Scholar
  23. Dhawan A, Vu CT, Zelikovsky A, Li Y, Prasad SK: Maximum lifetime of sensor networks with adjustable sensing range. Seventh ACIS International Conference on Software Engineering, Artificial Intelligence, Networking, and Parallel/Distributed Computing (SNPD'06) ed. by (Las Vegas, 2006), pp. 285–289Google Scholar
  24. Wang J, Medidi S: Energy efficient coverage with variable sensing radii in wireless sensor networks. Third IEEE International Conference on Wireless and Mobile Computing, Networking and Communications, 2007 ed. by (WiMOB 2007, White Plains, NY, 8–10 Oct. 2007), p. 61Google Scholar
  25. Lin S, Zang J, Zhou G, Gu L, He T, Stankovic JA: ATPC: adaptive transmission power control for wireless sensor networks. ACM Conference on Embedded Networked Sensor Systems (ACM SenSys’06) ed. by (Boulder, Colorado, 2006), pp. 223–236Google Scholar
  26. Rappaport TS: Wireless Communications: Principles and Practice. (Prentice Hall, Englewood Cliffs, NJ, 2002)Google Scholar
  27. Wolfram Research: Mathematica.[http://www.worlfram.com/]

Copyright

© Cheng et al.; licensee Springer. 2012

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.